inclusionAI/Ling-3.1-flash-base
just landedmodel
Ling-3.1-flash-base is the base checkpoint of Ling-3.1-flash, released to support research and community-driven innovation. Mixture-of-experts, 572B total, 25B active (512 experts, 8 per token). Context: 256K tokens (config). License: mit.
- Org
- inclusionAI
- Released
- 2026-10-25
- Status
- just landed
- License
- mit (permissive)
- Parameters
- 572.1B total, 25.0B active
- Architecture
- MoE · 512 experts, 8/token · BailingMoeV3ForCausalLM
- Context
- 262,144 tokens
- Task
- text-generation · text
- Signal
- 13 HF likes · 0 downloads
Runs on
mixed: real files + estimate; need = weights + FP16 KV cache for 8K tokens + 1GB. A sizing calculation, not tested.
- Q4 ~345.0 GB (estimated size): Mac 512GB
- Q8 ~609.6 GB (estimated size): no listed device
- FP16 ~1146.0 GB: no listed device
Links
First published here 2026-10-25 · updated 2026-10-27 · markdown · JSON